1. Field of the Invention
The present invention relates generally to optical beam steering and particularly to an optical system utilizing the electro-optical properties of liquid crystals to steer a light beam relative to a light-receiving end of an optical waveguide such as an optical fiber.
2. Description of the Related Art
N×N optical cross connect switches used, for example, in telecommunications systems, optically couple any one of N input optical fibers to any one of N output optical fibers. Such switches comprise examples of systems that make use of various techniques for deflecting or steering a light beam emitted from a light source, in this case the light-emitting end of a selected one of the N input fibers, to the light-receiving end of an optical waveguide, here in the form of a selected one (and in some cases, more than one) of the N output fibers.
One conventional optical beam steering technique applies diffractive beam steering that exploits the electro-optical properties of liquid crystals. This approach is analogous to the use of phased-array antennas for directing microwave radiation in radar systems and is the subject of a number of patents, for example, U.S. Pat. No. 5,093,747 to Dorschner and U.S. Pat. No. 5,963,682 to Dorschner, et al. These patents disclose a liquid crystal beam steering device comprising an optical beam phase shifting liquid crystal cell having a pair of spaced apart, parallel, superposed windows optically transparent at the wavelengths of interest. The pair of windows have inner, confronting surfaces. An electrically conductive, optically transparent (or reflective) common or ground plane electrode is affixed to the inner surface of one of the windows. A plurality of parallel, electrically conductive, optically transparent stripe electrodes are affixed to the internal surface of the other window. The space between the confronting electrode-bearing surfaces of the windows is filled with a layer of liquid crystal molecules, typically of the long, thin, rod-like organic type of the nematic phase. A periodic staircase waveform comprising voltage steps applied to the stripe electrodes creates corresponding local variations in the refractive index of the liquid crystal layer in such a manner as to form a diffraction grating of adjustable period. Thus, the wavefront of a light beam that emerges from the liquid crystal cell is tilted with respect to that of the incident wavefront. In this fashion, the optical beam phase shifting liquid crystal cell provides controllable beam steering as a function of the electrical potentials applied to the stripe electrodes. This conventional phased-array beam steering technique is digital in nature, with discrete voltages applied to the stripe electrodes (each representing a pixel) producing a stepped or staircase variation in the refractive index of the liquid crystal layer and a corresponding stepped or staircase variation in the optical phase delay during the transit of a light beam through the liquid crystal layer.
A drawback of striped electrode liquid crystal cell beam steering devices is that they tend to introduce crosstalk due to diffraction. Another disadvantage of striped electrode beam steerers systems is that they limit the available beam steering angles to discrete angular increments. This limitation results from the digital nature of these devices as well as from the striped electrode interconnection and drive schemes. In a conventional striped electrode liquid crystal cell, not all of the electrode elements are electrically independent; rather, every nth electrode is typically connected together to form periodically repeating electrode series or subarrays. The addressable beamsteering angles are restricted to those, that correspond to integer multiples of 360 degree (2π) phase ramps across each electrode subarray. Although large subarrays can accommodate many possible integer factors (and thus many steerable angles), the steerable angle is still limited to discrete increments; it is not continuously variable.
U.S. Pat. No. 4,852,962 discloses an optical fiber switch for switching light from one input fiber to any one of several output fibers. The switch includes a light deflection or steering cell consisting of two glass plates, nematic liquid crystal material between the glass plates, and spacers and electrodes formed on both sides of elongated electrode holders disposed along opposite edges of the glass plates, that is, along opposite sides of the light beam whose direction is to be controlled. Each electrode pair is controlled by an independent voltage source. The '962 patent states that when the intensity of the electric fields produced on the two sides of the beams differ, different orientations of the nematic crystal material occur across the beams, and depending on the difference in the field strength between the two electrode pairs the light beams are deflected through a larger or smaller angle. The deflections of the light is said to determine which one of the multiplicity of output fibers receives the light emitted form the single input fiber. A “continuous gradient” of refractive index is said to be formed across the light beams, but the '962 patent does not make clear how this is achieved, nor does the patent describe the distribution of the “gradient”, for example, whether it is linear.
Liquid crystal beam steering devices of the prior art deflect a light beam to couple a selected input optical fiber with a selected one (or more) output fiber(s). None of those devices deal with the problem of aligning the light beam relative to the light-receiving end(s) of the selected output fiber(s) to correct, for example, for the angle of the incident or emitted beam or for lens imperfections. In an N×N optical switch, the focal points of the light beams passing through the coupling lenses must precisely align with the receiving end surfaces of the output optical fibers to minimize coupling losses. A misalignment of even a few microns substantially reduces the optical coupling efficiency and, as N increases even modestly, the number of different combinations of connections (N!) becomes very large. Thus, to mechanically fix the relative positions of the fiber ends, microlenses and other optical elements so that each input fiber is properly aligned optically with all of the output fibers would be extremely difficult and costly.
FIG. 1 is a graph showing the rapid decrease in coupling efficiency as a function of fiber position error for a commonly used communication optical fiber, namely, a single mode fiber having a core diameter of 8 microns and a numerical aperture of 0.092. For example, it will be seen that for a fiber misalignment of 4.5 microns, the coupling efficiency is halved. Spatial registration precision of 1-2 microns and 1/10° of angular precision are required for adequate coupling efficiency.